Releases: LanguageMachines/frogdata
Releases · LanguageMachines/frogdata
v0.22
[Ko van der Sloot
- updated mwu data
- updated mblem data
v0.21
[Ko van der Sloot]
- Updated 'dum' for MBT version 2 (see: #7)
v0.20
- retrained all Mbt files (tagger, chunker, ner) with the new UTF8 aware Mbt
- retrained Timbl trees (Mblem and Mbma) with the new UTF8 aware Timbl
v0.18
[Ko van der Sloot]
- added some comment to nld/frog.cfg
- added some test files to test using external Mblem/MBMA and Mbt servers
- updated NER data
[Maarten van Gompel]
- added a nld-vnn configuration
v0.17
[Ko van der Sloot]
- small fix in mblem data, and retrained
- updated NER gazetteer files
- added cgn constraints file
[Maarten van Gompel]
- config/nld/ners.known, config/nld/plaatsen-nld.ner: added dutch
cities/villages, may overlap with geonames but this is more specific
- config/nld/ners.known, config/nld/straten-nld.ner: Added dutch
streets gazetteer data (extracted from OpenStreetMap)
- Added middle dutch model (INL/nederlab-linguistic-enrichment#15) in /dum/
subdir
v0.16
- cleanup and additions to the NER gazeteer files
v0.15
- new enhanced NER data, including gazeteer data
- new enhanced IOB data
- numerous fixes in MBMA rules
- some updates in MBLEM
v0.13
- Data files are now in
share/
instead of etc/
(incompatible with frog < 0.13.7)
v0.12.1
Minor bugfix release to facilitate debian packaging
v0.12
0.12 [Ko van der Sloot] 11-07-2016
- generally use ISO 639-3 for language codes
- quite a big update in the MBMA rules